AI Unicorns Surge as Venture Deal Count Hits 18-Month Low

Fewer bets, bigger winners

The third quarter of 2026 pulled AI venture capital in two directions at once. CB Insights reports that AI deal count fell to its lowest level in 18 months. Over the same period, 42 companies reached billion-dollar valuations, the highest quarterly total in a year and a half 1. Its summary is simple: investors are making fewer AI bets overall but are still minting new unicorns 1.

Crunchbase data, as summarized by OpenVC, points the same way from a dollars-raised angle. Global venture funding came to $159 billion for the quarter. North American funding dropped 35% to $92 billion as the very largest frontier-lab rounds, the kind associated with OpenAI and Anthropic, took a pause 2. AI still absorbed $102 billion, or about 64% of all global venture capital. Just 27 companies collected roughly a third of total funding through single checks of $1 billion or more 2.

Taken together, the two datasets describe a market that is shrinking in breadth while staying concentrated, and in some respects growing, at the top.

Where the sources agree, and where they differ in emphasis

Both datasets show capital flowing heavily into a small set of AI companies. CB Insights focuses on the count side: fewer transactions, more unicorn births 1. The Crunchbase figures focus on dollar concentration: a two-thirds AI share of global funding and a handful of mega-round recipients taking an outsized slice 2.

The framing differs. CB Insights' headline treats the unicorn surge as evidence that investor appetite for AI remains strong despite the slowdown 1. The Crunchbase-based analysis is more cautious. It presents the mega-rounds as coinciding with a widening funding gap for ordinary founders, especially those outside AI 2.

These readings don't contradict each other. They are two views of the same structural shift.

The squeeze below the top tier

The most telling figures sit underneath the headline totals. According to the Crunchbase data, the share of seed-stage companies that go on to raise a Series A has fallen from 50% to 38% 2. The typical time between those rounds has stretched to 616 days, well over a year and a half of runway a founder must survive before institutional follow-on money arrives 2.

Valuation gaps tell a similar story. The median pre-money valuation for a non-AI Series A stands at $42.4 million. AI Series A rounds carry a median of $78 million, about 84% higher 2. A non-AI company therefore raises its first major institutional round at a substantially lower price, if it raises one at all.

This helps reconcile the low deal count with the record unicorn births. When capital concentrates, fewer companies get funded, but those that do get funded at larger sizes and higher valuations. That pushes more of them past the billion-dollar mark faster. A rising unicorn tally and a shrinking deal count can be two outcomes of the same investor behavior.

Why it matters

For investors, the data suggests a flight to perceived quality and scale within AI. Backing a company that already looks like a category leader appears to feel safer than spreading money across many early bets. The pause in the largest frontier-model rounds also shows how sensitive quarterly totals have become to a few transactions. When two or three labs skip a quarter, North American funding can drop by a third 2.

For founders, the implications split along the AI line. AI startups with traction appear able to command premium valuations and, in some cases, quick unicorn status 12. Founders outside AI, or AI founders who can't show early momentum, face longer gaps between rounds and tougher conversion odds 2. The financing "valley" between seed and Series A is a real constraint on company formation, and it could shape which startups exist in two or three years.

The reading

The Q3 numbers do not show the AI boom cooling. They show it narrowing. Unicorn creation remains strong, and AI's share of venture capital has rarely been higher 12. But the breadth of the market, measured by deal count and seed-to-Series A conversion, is clearly weakening 12.

In this analyst's view, the 42 new unicorns are less a sign of broad optimism than a symptom of concentration. Investors are writing fewer, larger checks to a narrower set of companies they expect to win. If that pattern holds, headline funding totals will stay volatile and dependent on a few mega-rounds. The more lasting effect may fall on the many startups that never reach a Series A at all.